GUIDELINES FOR SITES WITH < 512K AVAILABLE DATA-ONLY TO DC PROMO

      I would like to share the GUIDELINES FOR SITES WITH 512K AVAILABLE DATA-ONLY CONNECTIONS and PROPOSE TO DC PROMO THIER LOCAL DCs. For sites which plan to install and promote (locally) an AD domain controller, a 512K available Data-only connection is the strong recommendation. And the only connection I am quite confident will succeed and not require additional support and effort.   If the site has a link lower than that, additional research needs to be done to ensure a smooth promotion and to minimize adverse business impact.

      My recommendation that we feel comfortable with. is 512K data only.  Anything under that, we may be able to try, but unfortunately, we cannot guarantee anything.  So it is up to the site to determine if they want to take on that risk.  For example, if a site’s circuit is going to be upgraded later anyway, they may want to wait.  Unless there's a business case that indicates a site cannot wait.

Considerations:
       Those at the site representing the business must understand and agree to the additional risk, should they elect to promote a DC over a <512K available data-only  link, which includes:

• Very very slow connections to external resources, including applications, internet, etc. for a week or more.  They should expect 3 weeks.
• A successful promotion at a similar site does not ensure other similar sites will be successful.  Because every site is different, and the databases increase daily.

Here are the rules:
       If a site MUST DC Promo over a smaller (than 512 available -  meaning part of the link isn’t dedicated to some other data stream like Mail, ERP, etc.-   data-only) or shared link,  the Server Infrastructure Team needs to talk to them to gather important information.

• It is critical that the design teams understand what is going over the connection, so we can make an informed decision.
• When promotions fail, it causes rework and could delay other sites.   Also, the rework always introduces some small amount of risk that something inadvertently corrupts the rest of the forest. 
• Once we all understand what is going over the <512K link, the customer agrees to the risk, and the Server Infrastructure Team design team feels it will not adversely impact others, we can OK the attempt at a DC promotion.  Again, the site will be expected to significantly reduce any other traffic going over the link during the promotion, and also during the SMS build. Traffic what should be taken into consideration, and significantly curtailed includes:
        • voice traffic
        • external application traffic (like intranet, ERP)
        • Internet traffic
        • Promotions must start the Fri just before a weekend to ensure the best throughput.
        • Promote the DC at a well-connected site, and then ship to the other site whenever possible.
• We can't make a lot of special allowances trying to make it work, If it promotes, it promotes.  And if it doesn't, it doesn't.  In many cases, we can try it, if the customers willing to take on the risk. And once all servers come up on site, we cannot be certain there will be no performance degradation when servers try to sync etc….  Again, there are too many variables. 

      We need to address and agree to the plan for <512K available data-only sites well in advance of their planned DC promotions.  That way, things can run as smoothly as possible.

Read More

Application Service Provider Checklist Examples

       The purpose of "Application Service Provider Checklist" is to obtain background information for those external vendors (3rd parties) that are currently providing or plan to provide external application hosting services for your business.

Items	Service Provider	Response
A1	Provide the name of the Application Service Provider (Outsourcer) and business address.
A2	Provide the name of the application to be hosted at the provider’s location.
A3	How long have you performed as or provided Application Service Provider (ASP) hosting services?
A4	How many applications do you provide hosting services for?
A5	How many customers do you currently support?
	How many customers do you support for the application your company is interested in (if you host more than one application)?
A6	Do you provide both shared and dedicated infrastructure (application, database, O/S) hosting options? a. How many customers utilize your shared infrastructure? b. How many customers utilize your dedicated infrastructure? c. Do you have separate database instances for your customers or do they share the same database? d. Is the application, web, and database on separate servers? e. How many application servers are used to host the application? f. How many database servers are used to host the database? g. For web-based environments, is the web server installed on the same server as the application? If no, how many web servers are used to support the application?
A7	What IT governance or security framework do you use for your control environment (COBIT, ISO17799, ISO 27002 internal policies and standards, etc.)?
A8	Do you have an internal and/or external audit function?
A9	Have you contracted with a 3rd party to provide an attestation of your control environment (i.e. SAS70 certified, BITS)? a. Please indicate the name and how often performed b. Note: for SAS70 please indicate - Type I, II
A10	Has or will a major acquisition (merger) occur in the next 6-12 months?
A11	What is your core business (expertise)?
Items	Network	Response
B1	Describe all end to end encryption methods currently supported (i.e.  SSL, HTTPS, VPN, IPSEC, SFTP) to securely transport data between you and your customers – include strength of cipher (i.e. 128 bit)?
B2	Describe all email encryption methods you currently support (i.e. TLS, PGP, etc.).
B3	Are strong authentication measures (i.e. two-factor authentication using RSA tokens or smartcards) used for remote access to your network or for remote administration of network devices (i.e. firewalls, routers, switches, IDS, etc) Note: userid/password is single factor
B4	Is redundancy and/or failover employed for critical devices such as firewalls, servers, load balancers, etc.?  Please provide detail.
B5	Are intrusion detection or intrusion prevention systems used? a. Network Based – where deployed b. Host Based – where deployed c. Application based – where deployed
B6	Please provide information about vulnerabilities assessments performed for your environment: a. List the type of assessments performed (penetration tests, network vulnerability scanning, etc.) b. Describe the scope of the assessments ( network perimeter, application assessment, etc.) c. How often are they performed? d. Are they performed by internal staff or external parties?
Items	Operations	Response
C1	Where is the primary processing facility (data center) located?
C2	Are any functions outsourced to a 3rd party (i.e. application development, system or network admin, data center)?  Please describe.
C3	Is access to the datacenter where the IT infrastructure resides controlled by you or by a 3rd party?
C4	Describe your process for keeping abreast of security threats for network devices, database, and operating system components?
C5	Do you have procedures in place for incident response, escalation and investigation?
C6	Is a formal change control process used to manage and track customer change requests and changes to the application, database, network and operating system components?
C7	Are security threats (events) for the application, database and operating system logged and reviewed regularly?  How often?
C8	Do you have separate development, test and production environments?
C9	Does the application reside in the same domain as the applications used to support your business? Does the application and its components reside on a separate VLAN from other applications?
C10	Is user access to the application controlled by the customer or the Application Service Provider (i.e. add/remove users, password management, assign roles, etc.)?
Items	Disaster Recovery	Response
D1	Do you have a documented Business Continuity and Disaster Recovery Plan to address short term and long term disruptions of service?
D2	Are the plans reviewed and tested at least annually?
D3	Describe customer involvement in the annual testing.
D4	Where is your alternate processing facility located?
D5	Is the alternate processing facility a hot-site or cold-site?  If other please explain.
D6	What type natural disasters are common in the region where the primary data center is located?

Read More

New Infrastructure Systems

       This article contains the processes for New Infrastructure Systems.  The information contained within this document applies to sites that range from newly acquired sites to established sites.  These sites may range in size from small sites (0-300 PCs), medium (301-500 PCs), or Large (501-1000 PCs up).


Refer to this article if:

• Your site is currently not on the planed, namespace, and you plan to install your company workstations or servers. (That entire process is described here.)
• You need to perform a subset of a new infrastructure deployment, such as:
      • Installing a new subnet for a site
      • Setting up DHCP at a site
      • Setting up DNS at a site
      • Installing Organizational Unit (OU) structure at a site
      • Setting up AD (Sites and Services ) for a site
      • Installing an AD Domain Controller at a site
      • Installing a software management services server at a site

Site Services Definitions:  Subnets, DNS, DHCP, OU’s
Subnet Design
The network team, who makes updates to the router for all sites should plan/design as is required for the following:

• IP readdressing
• Subnet mask

Note:  This design is to plan minimum of 15 weeks before the infrastructure needs to be in place (your scheduled IP readdress date, or server/desktop deployment date). 

New sites
For those deploying newly acquired sites, the following steps will be required:

• Design new IP Addresses
• Design DNS Entries
• Subnet Design
• Work on site design
• Procure hardware and installation – Routers, rack, servers, space, power requirements

Procure Hardware
       Determine Hardware Requirement Determine your site’s hardware requirements and place your order 15 weeks prior to Day 1 of deployments. Depending on your location and procurement processes, it can take anywhere between 2 weeks and 4 months to receive your hardware.

Server Builds
       The individual server teams can be taked control to finish the process, which takes 5-7 work days for each server.) for example

• AD Server and DC Promo – DCs need 5-7 work days (assumes already racked, and turned  on “burning in the HW” for a  minimum of 48 hours, and ready to begin DC Promo by day 1) for servers to be built.  After building out the AD Server and promoting the DC, set up the trust between CT and the resource domain. 
• Software Management Services Server - SMS needs 5-7 work days (assumes already racked, built to Brand and turned on “burning in the HW” for a minimum of 48 hrs by day 1).  Do not plan on deploying workstations until a minimum of 10 work days after the SMS server is complete to avoid workstation deployment delay. The SMS and Exchange servers must wait for the DC server to be completed.  Exchange is also dependent on SMS being available. 
• DHCP/DNS Build (Create/Delegate DHCP Scope) – NS Servers need 5-7 work days for completion. 
• Exchange Server - The Active Directory and SAN storage? will need to be in place prior to server installation. Mailboxes can be migrated as soon as servers and storage are in place.  Note that if Exchange will be at the site, then a DC must be at the site as well.
• Print Server -  Build up to brand 8 weeks prior to PC Deployment

Read More

Data Center Fire Sprinkler System

       This standard operating procedure template provides guidelines for Data Center Fire Sprinkler System, operating, inspecting, and maintaining the fire sprinkler system at your site(s). This will be achieved by a workforce including, but not limited to, contractors, vendor partners and employees who consistently apply safe work practices. Safe work practices, including emergency procedures in the event of disasters, support Operational Excellence. A safe and secure environment helps ensure the health and well being of all individuals (including workforce and visitors) as well as minimizes the impact of incidents that could affect business operations.

      To mitigate potential and preventable impacts to data center operations that could be caused by a fire going undetected through the application of industry standards and current best practices with the goal of ensuring 100% operability and reliability when called upon for service.

      Workforce members should assist each other in following the guidelines.

Input(s)

• The Facilities Team is responsible for maintaining an operational fire sprinkler system.
• See contact list for appropriate contacts.
• Any use of chemicals, cleaner, lubricants, etc. in support of data center operations requires a material safety data sheet be submitted and approved by the data center operations manager prior to being brought onto the site.
• The workforce is responsible for proper housekeeping practices, including the storage of tools and equipment during and after the work, cleanup, and waste disposal.
• All tools and equipment must be stored in a pre-approved area, or removed from the work site.
• All regulatory and company safe work practices must be followed where applicable, including but not limited to personal protective equipment, safety barriers lockout/tagout, and confined space entry.
• Any member of the workforce may stop any work in progress due to unsafe work methods, conditions causing the area to be unsafe, emergency, or for any other data center operational necessity. All unsafe work methods will be reported to the data center operations manager who will investigate the situation and take corrective action.
• Depending on organizational capability to perform the required work and/or maintenance, it may be necessary to contract with a qualified third party for some or all of the work described below. Note that many authorities having jurisdiction require that a licensed fire protection contractor perform at least some of the work on these systems.  For instance, sprinkler head inspections are typically done by facility staff, while the replacement of a pre-action valve is typically done by a licensed contractor. These requirements shall be verified for each site prior to implementing this standard operating procedure.

System Description
—Include Site System Information Here

• The fire sprinkler system consists of a fire alarm panel, pump panel, pumps, piping, valves, and sprinkler heads.
• Location—Indicate where the system is installed to include the risers, pumps, fire alarm panel, and other system components. Indicate if the system is installed under the raised floor, above the dropped ceiling or only at the rack level as applicable.
• Specifically describe the operation of the pump panel.  Define how to run the pumps at no flow, minimum flow, rated flow, and peak flow conditions.
• Indicate if the use of passwords is required.
• Indicate the location of the Operations and Maintenance manuals for further reference.
• Reference the Fire Alarm Panel and Detection standard operating procedure for the following:
   • Detection sequence leading to alarm and discharge
   • How an alarm or fault is indicated both at the panel and at the horns and strobes for normal, alarm and fault conditions.
   • Fire panel buttons, menus and diagnostic testing.
   • Silencing alarms, clearing/resetting the system.
   • Alarm notification sequence and expected response times. 

Outputs
       Data center manager shall ensure that a logbook is kept up to date with all inspection results and required actions.

Metrics

• Fire sprinkler system remains 100% available.
• No false alarms.

Read More

Data Center Access Policy and Guidelines

       This procedure provides Data Center Access Policy and Guidelines for the process of providing short term (visitor) and long term data center access. This will be achieved by a workforce (i.e. anyone conducting work including, but not limited to contractors, vendor partners, employees, etc.) who consistently apply safe work practices, including emergency procedures in the event of disasters, to meet your business needs.  A safe and secure environment will help to assure the health and well being of all individuals (including workforce and visitors) as well as minimize the impact of incidents that could affect business operations. 
      To mitigate potential and preventable impacts to facilities IT equipment caused  by unauthorized data center access through the application of industry standards, governmental regulations and current best practices with the goal of ensuring 100% operability and reliability when called upon for service.
 
       Workforce members should assist each other in following the guidelines.

Input(s)

• The Data Center Facilities Team Lead is responsible to ensure that all IT personnel adhere to the policy of applying for data center access.
• See contact list for appropriate contacts.
• Any member of the workforce may STOP any WORK in progress due to unsafe work methods, conditions causing the area to be unsafe, emergency, or for any other Data Center operational necessity.  All unsafe work methods will be reported to the Data Center Operations Team Lead who will investigate the situation and take corrective action

Physical Security

• Physical access to all computer rooms must be tightly controlled.  Doors must be locked at all times with only authorized personnel having access.
    - All employees, contractors, and visitors on company premises must wear identification tags at all times.
    - All visitors and vendors MUST be approved by the Data Center Facilities authorized approvers, visitor access added to their smart card (visitor card), and must be escorted while in the data center.  - Authorized personnel must not allow unknown or unauthorized individuals into restricted areas.  Unauthorized or unknown personnel not accompanied by authorized personnel, particularly in computer areas, must be challenged (in a tactful manner).  Personnel without a valid reason for being in the computer room must be escorted out of the computer room immediately and Security must be contacted.
    - Security in the Data Centers is the responsibility of the Data Center Facilities Team Lead.  The Data Center Facilities Team Lead will manage security by:
          ** Daily reviews and processing of your facilities Data Center Access
          ** Weekly reviews - Door Checks
          ** Quarterly review of all individuals with Data Center Access
          ** Yearly Documentation and Review

Outputs
        Data Center Facilities Team Lead shall ensure that records and documentation is kept with all checklist results, door checks and required maintenance actions.


Metrics
       Check list results are maintained and available at all times.


Contacts

• Data Center Facilities Team Lead – xxx who will be assigned
• Data Center Facilities Coordinator – xxx who will be assinged
• Emergency Back Up – xxx who will be assigned

Read More

Best Practices of RAISED FLOOR

       The purpose of this article is to provide a basis of "Best Practices of RAISED FLOOR" for data center management and infrastructure. This can be applied to your data center. Insuring that the raised floor is structurally sound, well grounded and maintained in a proper manner contributes not only to the overall reliability of the data center; but, to safety as well.

Design
      The following should be considered and implemented in the design of the data center raised floor.

• The raised floor grid should be grounded to the ground reference (meaning earth ground).
• Use "white space" to spread out the equipment and prevent hot spots
• A floor tile layout grid marked either on the walls or the tiles themselves will provide an easy reference to any equipment location. This grid layout can also be indicated on any columns
• Any sub-floor infrastructure (valves, electrical panels, etc.) should be noted with signage. The same applies to any infrastructure in the ceiling that cannot be seen
• Fiber cabling runs in the raised floor should be protected in either metal cable raceways or by some other  protection method
• Servers should be elevated off of the floor so that air intakes do not become clogged with dust and dirt
• In areas where earthquakes are prevalent, earthquake restraints should be installed on all equipment, storage cabinets and shelving
• In earthquake prone areas, equipment should be installed on the Iso-Base product.
• All equipment racks should be equipped with a power distribution unit. The use of power strips to connect to electrical circuits is viewed as a safety hazard.

Operations
      The following should be considered and implemented in the operations of the data center raised floor.

• A proactive maintenance program is critical to long-term effectiveness of any program designed for system availability.
• A data center audit should be conducted on a regular basis by an outside, impartial firm
• The raised floor and pedestals should be checked on a regular basis for leveling and fit. Warped, protruding or badly fitting tiles should be replaced.
• The sub-floor are should be cleaned with a vacuum on a regular basis in order to prevent debris from being blown into the equipment racks.
• Tiles that have galvanized, non-painted surfaces on the bottom of the tiles should be checked for zinc whiskers.  The zinc in galvanized tiles will come off in 2-micron whiskers. These may be blown into the equipment racks.
• Unused data cables should be removed from the sub-floor. This is a NFPA code requirement.
• Marking tiles or using contrasting color tiles to indicate where equipment can be installed versus reserved space for infrastructure or white space alleviates layout confusion
• Use yellow hazard tape on the tiles to guide foot traffic away from critical or non workspace areas
• Raised flooring that has electroplated passivated sheet metal bottoms with wood cores is subject to zinc whiskers. These are approximately 2 microns thick, and can be blown into equipment. Inspect the bottoms of the tiles with a flashlight at an oblique angle. If the surface twinkles, zinc whiskers are present.
• At no time should wire spools with excess wire or cable be left under the floor. These contribute to poor air distribution
• Equipment should be unpacked outside the data center to reduce the amount of debris and dirt circulating in the space
• Storage of cardboard boxes in the data center should not be permitted. Doing so is a fire hazard and contributes to the amount of dust and fiber circulated into the equipment.
• No food or drink should be allowed in the data centers at any time.

Read More

Infrastructure Equipment

       The purpose of this article is to provide a basis of "Best Practices of Infrastructure Equipment" for data center management and infrastructure. Local business reasons, local governmental code or other circumstances may mitigate the implementation of these best practices.


       It should be understood that the implementation of "Best Practices of Infrastructure Equipment" may require long term processes to be implemented, capital or expense funding and that the corrective actions will be ongoing. No work activities can be performed that have the potential to impact services without the prior approval of the appropriate management personnel.

• Crankcase vapor recover systems on the diesel generators help keep the fins clean and heat transfer efficient
• Fuel filters on the diesel generators should have water detection sensors installed
• Having multiple fuel and oil filters with bypass valving on the diesel generators allows filter replacement while the generators are running
• Re-circulation dampers reduce CFM air flow during cold weather
• The installation of an insulating jacket of the diesel generator exhaust pipes will help to reduce ambient temperatures inside enclosed engine rooms
• Compressed air should be piped directly to the diesel generator locations, with the proper quick connect fittings, to provide easy connection points for impact tools
• The use of aircraft type of hose (metal braided exterior) vs. the rubber hoses on diesel generators will provide lengthened hose life
• Diesel generator crankcase heaters should have cutoff valves installed for easy hose replacement
• Pre-filled oil filter cartridges and an electric lube oil transfer pump can reduce downtime for an oil change
• Engine start batteries should have clear plastic covers installed to prevent accidents
• Permanently installed load banks will allow UPS systems and generators to be tested on a regular basis. Both UPS systems and generators should be tested on a monthly basis
• A log sheet for recording UPS maintenance should be readily available, and preferably attached to the main UPS panels
• All circuit breakers should be marked with a color coded schema to indicate whether they are normally open or closed
• Main UPS and switch gear circuit breakers should be protected from accidental operation
• Spare fuses should be stocked near the points where they are used
• Environmental equipment systems labeling should be thorough, consistent and clear
• Acid absorbent materials should be installed below UPS batteries
• The installation of a permanent chain host in the UPS battery room will facilitate battery replacement and reduce the chance of personnel injury
• Individual cell equalizers installed on the UPS batteries will ensure that each cell gets the exact charging voltage required for optimal battery performance
• The use of a battery watering cart, complete with a bulk supply of de-ionized water and a small pump, will reduce the time required for topping off batteries
• If battery lugs have multiple bolt holes, multiple bolts should be used to prevent constraint of current flow
• Eye wash stations in the battery rooms should be interconnected to the alarm system, with appropriate organizations notified when the station is used
• All piping should be numbered, color coded and flow direction noted
• External connection points on the chilled water piping would allow the use of truck mounted chiller units in case of emergency
• Breathing apparatus should be located in the proper area for the number of personnel assigned to the area
• The condenser and cooling tower water control valves should have a manual override with a valve position indicator clearly marked
• Cooling tower sump water should be checked on a daily basis for clearness. The use of a portable sand filter for the cooling towers is advantageous
• An active biocide program should be initiated to prevent growth in the condensate drains for the CRAHs
• Spill containment equipment and programs should be instituted in all spaces
• Having an abnormal operating configurations or conditions posted on a status board available to all shifts will allow each shift to immediately be informed about any conditions
• Each door exiting the equipment plants should have a telephone, flashlight, fire extinguisher and emergency procedures
• Lockout and tag out tools are clearly labeled and stored in such a manner that any missing tool is apparent
• Spare parts cabinets should be well stocked. Parts inventory sheets direct mechanics and engineers to the appropriate spares quickly
• Fire stopping between spaces is consistent and well done
• Fire detection heads should be on flexible conduit for easy relocation
• Cooling condenser coils should be kept clean. A regular maintenance program should exist on the coils. Dirty coils reduce capacity
• Cooling tower piping should be protected against freezing on cold days

Read More

Best Practices Data Center Safety

       The purpose of this article is to provide a basis of "Best Practices Data Center Safety" for data center management and infrastructure. Local business reasons, local governmental code or other circumstances may mitigate the implementation of Data Center Electrical. This can be applied to your business.


General development for good Back-protection habits:

• Exercise regularly for strong, flexible muscles
• Avoid or control stress
• Do not overestimate your own strength
• Walk, do not run, to prevent slips and falls
• Sleep on a firm mattress, either on your side with your knees bent or on your back with your knees elevated

Seek Alternatives to lifting:

• Use equipment: hoists, hand trucks or dollies, rather that lifting when possible
• Push, do not pull, loaded trucks or dollies
• Break a large load into several small ones
• Get a helper to lift heavy or awkward objects
• Plan jobs and work areas to minimize the need to move tools and materials

Before you lift:

• Plan straightest, fastest and clearest route to your destination
• Look for places to stop and rest along the way
• Remove any objects you may trip over before starting to move the object
• Try to have a waist high surface for loading
• Make sure the area where you will unload is clear
• Check the object you will be carrying for rough or jagged edges or slippery surfaces
• Lift properly, a corner of the object to check weight and stability
• Wear gloves with good grip, safety shoes with reinforced toes and nonskid soles, and clothing that fits snugly (may just want to limit this to gloves but to be aware of the others as a risk possibility)
• Warm up with gentle bends and stretches

Lift correctly:

• Stand close to load, with feet firmly on the floor, about shoulder width apart, with toes pointing in an outward position
• Squat down close to the load with back straight, knees bent, stomach muscles tight
• Place hands on the diagonally opposite corner of the load so one hand pulls the load toward you and the other hand lifts
• Grip load firmly with both hands, not just the fingers
• Bring the load as close as possible to your body, with weight centered over you feet, lock arms and elbows into side and chin into neck
• Stand up slowly, keep your back straight and let your legs do the lifting
• Move forward with small steps with the load close to the body and not above waist high
• Change direction by moving feet not by twisting

Unload properly:

• Lower body slowly, with knees bent so legs do the work
• Position hands so fingers will not be caught under the load
• Place load on edge of surface and slide it back
• Fire suppression systems should be tested twice yearly at a minimum for proper operation. Local codes may require more frequent testing.
• Fire extinguishers should be checked monthly for proper certification and levels.
• No equipment may be moved without data center management permission.
• Workstation areas (tables, desks, etc) that are used on more than an intermittent basis should comply with CITC ergonomic standards.
• No data cables should be run across aisle ways, including extension cables or power cables.
• Alarm procedures should be reviewed on an annual basis for compliance with local and national codes; and, adherence to your business safety policies.
• The location of all fire extinguishers should be clearly marked
• The location of all safety equipment should be clearly marked
• The location and operation of all EPO's, fire suppression and alarm switches should be clearly marked
• Each telephone in the data center should have an up to date contact list and emergency procedures listing attached nearby
• Each door exiting the data center should have a telephone, flashlight, floor puller, fire extinguisher and emergency procedures
• Each major piece of infrastructure equipment (PDU/CPC,  air handlers) should have an emergency contact number attached to it
• Fire stopping between spaces is consistent and well done
• Fire detection heads are on flexible conduit for easy relocation
• The following activities have been restricted from inside the Data Center, except as specifically authorized by the your IT manager

Read More

Best Practices Data Center Electrical

      The purpose of this article is to provide a basis of "Best Practices Data Center Electrical" for data center management and infrastructure. Local business reasons, local governmental code or other circumstances may mitigate the implementation of Data Center Electrical.
       It should be understood that the implementation of best practices Data Center Electrical may require long term processes to be implemented, capital or expense funding and that the corrective actions will be ongoing.

• Two electricians should be required for all “hot” electrical work.
• Infrared inspections of all PDU/CPCs should be performed yearly. Circuits that indicate readings 10 – 20% more than the norm should be noted and investigated.
• Unused circuits should be turned off at the PDU/CPCs.
• All PDU/CPCs should have a ground strap connected to the raised floor or ground.
  •Harmonic distortion, both voltage and current, measurements should be performed yearly or when substantial amounts of new equipment are installed; or major changes in the data center occur.
• The signal reference grid should be checked for integrity and measured ground resistance on a yearly basis.
• The proper labeling of PDU/CPC circuit breakers and circuits should be checked on a yearly basis. Panel schedules should be updated immediately when circuits are added or de-installed
• Loads on PDU/CPCs should be evenly distributed across the phases in order to reduce harmonics.
• An independent ground wire, should be run with each circuit. Circuits should never share ground wires.
• An independent neutral wire should be run with each circuit. Circuits should never share neutrals. The sizing of the neutral conductor depends on the load and the harmonics present.
• Watertight receptacles should be used beneath the sub-floor if chilled water air handlers are installed in the space.
• Use of 'pin and sleeve' receptacles should be used. If not available, or used, twist-lock receptacles should be raised off the floor by approximately 1 5/8".
• Branch circuits should be designed with the following factors: 1200 VA on 15 Amp, 120 Volt circuits; 1800 VA on 20 Amp, 120 Volt circuits and 3600 VA on 30 Amp, 120 Volt circuits. The minimum conduit size should be 1/2". Individual neutral and ground conductor in each computer receptacle branch circuit with a 3% maximum voltage drop at the furthest outlet in branch circuits and a 5% maximum voltage drop from the main power source to the load.
• The grounding scheme design should provide equal ground potential for all metal supports, equipment enclosures, racks, raceways and raised floor.
• The IEEE Emerald Book should be the standard for grounding all facilities (signal reference grid).
• Unused Seal-Tight circuits should be removed from the sub-floor. This is NFPA code requirement; and, it helps to insure PDU/CPC panel schedules are kept current. Additionally, unused circuits impede proper airflow.
• When using standard 110VAC NEMA 5 plugs, tie wrap the plugs to the receptacles to insure that the plugs do not come loose. Some locations may choose not to allow these circuits under the floor in order to prevent unauthorized personnel from utilizing those circuits.
• All circuit breakers should be marked with a color coded schema to indicate whether they are normally open or closed. This includes PDU/CPC mains and branch circuits.
• Manual EPO switches should be setup as a multiple step device and protected with a cover. If the cover is lifted, an alarm is sounded indicating imminent activation. The EPO switch should have a back-lit and a printed warning attached. The location and layout of the EPO switches should be consistent for each location in order to reduce confusion. Note, check with the manufacturer of the EPO system and relevant UL & NRTL code before any modifications to the EPO system.
• UPS power in the data center will only be used for computer and communications equipment. At no time will the UPS power be used for test, construction, personal and janitorial equipment
• At no time will a IT workforce member operate a PDU or main circuit breaker without consulting office management personnel.

Read More

Best practices Data Center Temperature

     The purpose of this article is to provide a basis of "Best practices Data Center Temperature" for data center management and infrastructure. Local business reasons, local governmental code or other circumstances may mitigate the implementation of these best practices.

Air Conditioning & Cooling (Air Handlers, CRAH or FCU units)

• Air handers should be equipped with filters. These filters should be checked weekly and changed as needed. Pleated filters should be equipped with pre-filters.
• The relative humidity should be set in the 55 – 60% RH range.
• Air handlers should be equipped with output thermostats. These thermostats should be tied into a master monitored console with trip point setup to activate an alarm.
• Air handlers should not be installed back to back. The resulting vortex effect on the discharge may actually reduce flow.
• Air handlers should not be located close to high heat load equipment. The static pressure in front of the air handlers is actually very low due to the high velocity of the air handler's discharge. The low static pressure may cause hot air above the raised floor to be sucked into the sub-floor. This problem may be pronounced with high-static-output air handlers.
• Air hander placement and the number of air handlers should be such that if any one air handler fails, the remaining units will provide sufficient pressure and flow to cool the equipment.
• The room's humidity and cooling should be checked on a semi-annual basis with independent measuring equipment for data centers that do not experience a high rate of change. For those data centers that experience constant change, this time frame should shrink.
• Any sub-floor impediments to proper airflow should be reviewed. Old cabling should be removed, and large cable bundles should be spread out.
• Air handlers should be checked on a regular basis for proper operation and unit performance. Sensors should be check against independent measurement devices.
• Spot temperature and relative humidity measurements should be made on a regular basis with calibrated independent instruments.
• The height of the raised floor contributes to adequate airflow. This coupled with the static output pressure from the air handlers will determine where physical hot spots may arise in the data center. With a two foot raised floor, under ideal conditions, a typical air handler will have enough static pressure for normal flow at 30 feet.
• The residual chilled water piping stubs that remain after de-installation of air handlers should be removed.
• The design of the equipment racks will contribute to adequate airflow. For racks that contain high heat loads, fully perforated front doors, or no doors, are preferential.
• Air handlers should be checked for excessive dehumidification caused by low coil temperatures. Excessive dehumidification will cause corrosion in the fan coil units. This can readily be checked if the coil units are wet
• Condensate drains under the air handlers should not drain into open floor drains
• Water piping for the humidifiers should be supported so that unintentional breakage of the lines do not occur
• The discharge area of the air handers should be kept clean so that dust, dirt and debris are not blown into equipment intakes
• An active biocide program should be initiated to prevent growth in the condensate drains for the CRAHs

Read More